The classification of the chain-link and other data by spherical SOM

A new classification method which can visualize a multidimensional data set was proposed. There, phase distances on the spherical surface among the label data was computed. Using these distances, a dendrogram can be constructed. Here, using the constructed dendrogram, the colored classification of the cluster group on the spherical surface based on the label data was carried out. This method can be applied to various data. Here, the method was applied to the chain-link problem which can be considered as difficult for the representation and to the problem of the separation of the random number planes which were piled up to the direction of the height.

Introduction

The new classification method which visualizes a multidimensional data set was proposed [1]. There, a phase distance on the spherical surface of the label data was computed and then, a dendrogram was constructed by the distance calculation among the labels. Here, the classification of the cluster group of the label data on the spherical surface [2], [3] was carried out based on this created dendrogram.